KIT | KIT-Bibliothek | Impressum | Datenschutz

NoRBERT: Transfer Learning for Requirements Classification

Hey, Tobias ORCID iD icon; Keim, Jan ORCID iD icon; Koziolek, Anne ORCID iD icon; Tichy, Walter F. ORCID iD icon


Classifying requirements is crucial for automatically handling natural language requirements. The performance of existing automatic classification approaches diminishes when applied to unseen projects because requirements usually vary in wording and style. The main problem is poor generalization. We propose NoRBERT that fine-tunes BERT, a language model that has proven useful for transfer learning. We apply our approach to different tasks in the domain of requirements classification. We achieve similar or better results (F$_1$-scores of up to 94%) on both seen and unseen projects for classifying functional and non-functional requirements on the PROMISE NFR dataset. NoRBERT outperforms recent approaches at classifying nonfunctional requirements subclasses. The most frequent classes are classified with an average F$_1$-score of 87%. In an unseen project setup on a relabeled PROMISE NFR dataset, our approach achieves an improvement of ten percentage points in average F$_1$- score compared to recent approaches. Additionally, we propose to classify functional requirements according to the included concerns, i.e., function, data, and behavior. ... mehr

Volltext §
DOI: 10.5445/IR/1000150464
Veröffentlicht am 08.09.2022
Cover der Publikation
Zugehörige Institution(en) am KIT Institut für Programmstrukturen und Datenorganisation (IPD)
Publikationstyp Forschungsbericht/Preprint
Publikationsjahr 2022
Sprache Englisch
Identifikator KITopen-ID: 1000150464
Verlag Karlsruher Institut für Technologie (KIT)
Bemerkung zur Veröffentlichung Corrected Version due to a bug in a part of the evaluation. Changes to original version are marked with red changebars in full text PDF.
Relationen in KITopen
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page